Transferability of neural networks approaches for low-rate energy disaggregation

David Murray, Lina Stankovic, Vladimir Stankovic, Srdjan Lulic, Srdjan Sladojevic

Research output: Contribution to conferencePaper

Abstract

Energy disaggregation of appliances using non-intrusive load monitoring (NILM) represents a set of signal and information processing methods used for appliance-level information extraction out of a meter's total or aggregate load. Large-scale deployments of smart meters worldwide and the availability of large amounts of data, motivates the shift from traditional source separation and Hidden Markov Model-based NILM towards data-driven NILM methods. Furthermore, we address the potential for scalable NILM roll-out by tackling disaggregation complexity as well as disaggregation on houses which have not been 'seen' before by the network, e.g., during training. In this paper, we focus on low rate NILM (with active power meter measurements sampled between 1-60 seconds) and present two different neural network architectures, one, based on convolutional neural network, and another based on gated recurrent unit, both of which classify the state and estimate the average power consumption of targeted appliances. Our proposed designs are driven by the need to have a well-trained generalised network which would be able to produce accurate results on a house that is not present in the training set, i.e., transferability. Performance results of the designed networks show excellent generalization ability and improvement compared to the state of the art.

Conference

Conference2019 International Conference on Acoustics, Speech, and Signal Processing
Abbreviated titleICASSP 2019
CountryUnited Kingdom
CityBrighton
Period12/05/1917/05/19

Fingerprint

Neural networks
Monitoring
Smart meters
Source separation
Hidden Markov models
Network architecture
Signal processing
Electric power utilization
Availability

Keywords

  • energy analytics
  • non-intrusive load monitoring
  • energy disaggregation
  • neural networks
  • deep learning

Cite this

Murray, D., Stankovic, L., Stankovic, V., Lulic, S., & Sladojevic, S. (Accepted/In press). Transferability of neural networks approaches for low-rate energy disaggregation. Paper presented at 2019 International Conference on Acoustics, Speech, and Signal Processing, Brighton, United Kingdom.
Murray, David ; Stankovic, Lina ; Stankovic, Vladimir ; Lulic, Srdjan ; Sladojevic, Srdjan. / Transferability of neural networks approaches for low-rate energy disaggregation. Paper presented at 2019 International Conference on Acoustics, Speech, and Signal Processing, Brighton, United Kingdom.5 p.
@conference{61fa9334e1ee47e6b1cd111c74d062c1,
title = "Transferability of neural networks approaches for low-rate energy disaggregation",
abstract = "Energy disaggregation of appliances using non-intrusive load monitoring (NILM) represents a set of signal and information processing methods used for appliance-level information extraction out of a meter's total or aggregate load. Large-scale deployments of smart meters worldwide and the availability of large amounts of data, motivates the shift from traditional source separation and Hidden Markov Model-based NILM towards data-driven NILM methods. Furthermore, we address the potential for scalable NILM roll-out by tackling disaggregation complexity as well as disaggregation on houses which have not been 'seen' before by the network, e.g., during training. In this paper, we focus on low rate NILM (with active power meter measurements sampled between 1-60 seconds) and present two different neural network architectures, one, based on convolutional neural network, and another based on gated recurrent unit, both of which classify the state and estimate the average power consumption of targeted appliances. Our proposed designs are driven by the need to have a well-trained generalised network which would be able to produce accurate results on a house that is not present in the training set, i.e., transferability. Performance results of the designed networks show excellent generalization ability and improvement compared to the state of the art.",
keywords = "energy analytics, non-intrusive load monitoring, energy disaggregation, neural networks, deep learning",
author = "David Murray and Lina Stankovic and Vladimir Stankovic and Srdjan Lulic and Srdjan Sladojevic",
note = "{\circledC} 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.; 2019 International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 ; Conference date: 12-05-2019 Through 17-05-2019",
year = "2019",
month = "2",
day = "1",
language = "English",

}

Murray, D, Stankovic, L, Stankovic, V, Lulic, S & Sladojevic, S 2019, 'Transferability of neural networks approaches for low-rate energy disaggregation' Paper presented at 2019 International Conference on Acoustics, Speech, and Signal Processing, Brighton, United Kingdom, 12/05/19 - 17/05/19, .

Transferability of neural networks approaches for low-rate energy disaggregation. / Murray, David; Stankovic, Lina; Stankovic, Vladimir; Lulic, Srdjan; Sladojevic, Srdjan.

2019. Paper presented at 2019 International Conference on Acoustics, Speech, and Signal Processing, Brighton, United Kingdom.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Transferability of neural networks approaches for low-rate energy disaggregation

AU - Murray, David

AU - Stankovic, Lina

AU - Stankovic, Vladimir

AU - Lulic, Srdjan

AU - Sladojevic, Srdjan

N1 - © 2018 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Energy disaggregation of appliances using non-intrusive load monitoring (NILM) represents a set of signal and information processing methods used for appliance-level information extraction out of a meter's total or aggregate load. Large-scale deployments of smart meters worldwide and the availability of large amounts of data, motivates the shift from traditional source separation and Hidden Markov Model-based NILM towards data-driven NILM methods. Furthermore, we address the potential for scalable NILM roll-out by tackling disaggregation complexity as well as disaggregation on houses which have not been 'seen' before by the network, e.g., during training. In this paper, we focus on low rate NILM (with active power meter measurements sampled between 1-60 seconds) and present two different neural network architectures, one, based on convolutional neural network, and another based on gated recurrent unit, both of which classify the state and estimate the average power consumption of targeted appliances. Our proposed designs are driven by the need to have a well-trained generalised network which would be able to produce accurate results on a house that is not present in the training set, i.e., transferability. Performance results of the designed networks show excellent generalization ability and improvement compared to the state of the art.

AB - Energy disaggregation of appliances using non-intrusive load monitoring (NILM) represents a set of signal and information processing methods used for appliance-level information extraction out of a meter's total or aggregate load. Large-scale deployments of smart meters worldwide and the availability of large amounts of data, motivates the shift from traditional source separation and Hidden Markov Model-based NILM towards data-driven NILM methods. Furthermore, we address the potential for scalable NILM roll-out by tackling disaggregation complexity as well as disaggregation on houses which have not been 'seen' before by the network, e.g., during training. In this paper, we focus on low rate NILM (with active power meter measurements sampled between 1-60 seconds) and present two different neural network architectures, one, based on convolutional neural network, and another based on gated recurrent unit, both of which classify the state and estimate the average power consumption of targeted appliances. Our proposed designs are driven by the need to have a well-trained generalised network which would be able to produce accurate results on a house that is not present in the training set, i.e., transferability. Performance results of the designed networks show excellent generalization ability and improvement compared to the state of the art.

KW - energy analytics

KW - non-intrusive load monitoring

KW - energy disaggregation

KW - neural networks

KW - deep learning

M3 - Paper

ER -

Murray D, Stankovic L, Stankovic V, Lulic S, Sladojevic S. Transferability of neural networks approaches for low-rate energy disaggregation. 2019. Paper presented at 2019 International Conference on Acoustics, Speech, and Signal Processing, Brighton, United Kingdom.